Computer graphics processing and selective visual display system – Display peripheral interface input device
Reexamination Certificate
1998-01-16
2001-02-27
Brier, Jeffery (Department: 2672)
Computer graphics processing and selective visual display system
Display peripheral interface input device
C345S156000, C345S182000
Reexamination Certificate
active
06195081
ABSTRACT:
BACKGROUND
The present disclosure relates generally to computers, and more specifically, to a system and method for reducing a resolution of a graphical image for display on a computer.
There are many different types of display devices requiring different formats of data to display. For example, a 640×480 pixel display, with each pixel having one of 256 colors selected from a larger group of 262,144 possible colors, is commonly used in personal computers. However, other displays, or other display drivers such as ones that do not use a VESA standard video BIOS or cannot support 256 colors, instead may only utilize 16 colors. As a result, any 256 color images must be converted to 16 colors.
Previous methods to display images on a reduced color set, such as a 256 color “source” image to a 16 color “target” image, include dithering and color quantizing. Conventional dithering involves interspersing colors to fool the eye into seeing a third, or intermediate, color. While dithering usually supplies more color precision, it is at a cost of image detail. Such image detail can be critical. For example, in graphic images including a combination of text and background, if image detail is lost, the text blends-in to the background and becomes difficult to read.
Conventional color quantization is a process of reducing the source images number of colors by merging similar colors together, until the target number of colors is reached. It requires extensive processing, performing multiple passes through the source image to gather information about the frequency and variety of colors used to determine an optimal color set for the target image. Once created, the colors can be modeled as a three dimensional space, such as one defined by a red axis, a green axis and a blue axis (RGB color space). In the RGB color space, all colors exist, with black being at the origin (0,0,0) and white being the furthest point away from the origin. Once the color space is created, it is divided into sub-areas, in which each sub-area (containing many colors) maps to a single target color. In this way, many different colors of the source image are mapped to a single target color. As a result of such extensive processing, quantization error for the colors of the source image are minimized.
Although color precision is lost in color quantization, image details are preserved. In the example graphic image having a combination of text and background as discussed above, with color quantization, image detail is retained and the text does not blend-in to the background. However, since color precision is lost, the text can become difficult to read if the color of the text and the color of the background are reduced to similar colors.
What is desired, therefore, is a reduction process that is quick, i.e., does not require extensive processing. Furthermore, what is desired is a reduction process that produces a target image with accurate image details. Further still, what is desired is a reduction process that makes text easy to read against a background despite the reduction in colors.
SUMMARY
In response thereto, provided is a system and method for performing a single-pass color quantization for graphic images on a computer. In one embodiment, the computer includes a processor capable of running operating code, an input device accessible by the processor for providing a source image, and an output device for displaying a target image. Both the source image and target image include a plurality of pixels, each having red-green-blue components. However, there are fewer available colors for the target image than for the source image.
A program instructs the processor to reduce the source image to the target image by reading a pixel from the source image and comparing the color of the pixel to mappings stored in a table. If the pixel color does not correspond to a mapping in the table, then the program instructs the processor to convert the pixel color into hue-whiteness-blackness (HWB) components, map the HWB components to a color of the target image, and store the mapping in the table. If the pixel color does correspond to a mapping in the table, then the program instructs the processor to simply retrieve the mapping.
In another embodiment, the program de-emphasizes the hue component by reducing its value before mapping it to a color of the target image.
In yet another embodiment, the program provides the mapped color to the output device for displaying the second image.
REFERENCES:
patent: 4823120 (1989-04-01), Thompson et al.
patent: 5333243 (1994-07-01), Best et al.
patent: 5650942 (1997-07-01), Granger
patent: 5930387 (1999-07-01), Chan et al.
patent: 6014133 (2000-01-01), Yamakado et al.
patent: 6016360 (2000-01-01), Nguyen et al.
Alvy Ray Smith & Eric Ray Lyons, HWB—A More Intuitive Hue-Based Color Model, pp. 1, and 1-2.
Blackman Anthony
Brier Jeffery
Dell USA L.P.
Haynes and Boone LLP
LandOfFree
Single-pass color quantization for graphic images does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Single-pass color quantization for graphic images, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Single-pass color quantization for graphic images will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2609930